Brain metastases (BrM) refers to a disease in which tumor cells originating in other parts of the body metastasis into the skull, causing tumors to grow in the skull, with lung cancer brain metastasis being the most common, followed by breast cancer and melanoma.
According to statistics, the probability of brain metastasis in cancer patients is as high as 20% to 40% [1], and even with aggressive combination of treatment, the median survival rate of 2-year and 5-year brain metastase patients is only 8.1% and 2.4% [2].
Unfortunately, little is known about the molecular mechanisms by which tumor cells metastasize and colonize the brain, and the lack of detailed information on the composition and functional status of cells within brain metastases has prevented researchers from conducting in-depth research on brain metastases.
Recently, Professor Hugo Gonzalez's team from the University of California, San Francisco, through single-cell sequencing (scRNA-seq) of brain metastase tissues and combining techniques such as mass spectrometry flow (CyTOF), revealed that brain metastase cells take proliferation and inflammation as the main two cellular states, and summarized and analyzed the relationship between immune cells in the matrix and the cellular microenvironment, which helped to understand the characteristics of brain metastase cells and their interaction with the cell microenvironment, and the relevant results were published in " Cell Journal[3].
▲Screenshot of the first page of the paper
The researchers collected brain metastase specimens from 15 patients with primary tumors such as melanoma (n=3), breast cancer (n=3), lung cancer (n=3), ovarian cancer (n=2), colorectal cancer (n=1), kidney cancer (n=1), adult rhabdomyosarcoma (n=1), and unknown primary cancer (n=1), sequenced them single-cell, and performed mass spectrometry flow analysis based on immune phenotypes in 11 of the samples.
A total of 80,377 single-cell transcripts were captured, and the cells were annotated by evaluating marker expression, chromosomal aberrations, copy number variations, etc., marking them as 49,488 brain metastase cells (MTC) and 30,889 brain metastase-associated non-malignant stromal cells.
Interestingly, tumor stromal cells from different patients form multiple clustered clusters in the dimensionality reduction graph (UMAP), while brain metastase cells in each patient form independent clusters with no overlap between each case, suggesting that brain metastase cells have a high degree of inter-patient heterogeneity (Figure 1).
Figure 1 Perform scRNA-seq and CyTOF on brain metastase tissues of 15 patients, reduce dimensions by UMAP, and annotate the cells
The researchers first analyzed the stromal cells in the tumor and found that they consisted mainly of vascular cells (endothelial and parietal cells), inflammatory immune cells, and mesenchymal progenitor cells, and identified 20 different clusters based on the expression of the marker gene (Figure 2).
Stromal cells in Figure 2 tumors: endothelial cells with the marker genes CLDN5 and PECAM1 (EC-1, EC-2 and EC-3 clusters), parietal cells (PC) and vascular smooth muscle cells (vSMCs, PC-1, PC-2, PC-3 clusters) with the marker genes ISLR and CTHRC1, mesenchymal stromal cell-like cells (MSC-like-1 and MSC-like-2), T cells with the marker genes CD3D and IL7R (T:CD8+:EM, T:CD4+:CM1, T:CD4+:cm2, Tregs and T:Cm), B cells with the marker genes JCHAIN and MZB1 (B-c1 and B-c2), transfer-associated macrophages with the marker genes AIF1 and LYZ (MAMs:APOE+ and MAMs:S100A8+), Dendritic cells with the marker genes CD1C and CLEC10A (DC, cDC2:CD1C+/CLEC10A+) and astrocytes with the marker genes GFAP and S100B
Blood vessel cells in the matrix make up the blood-tumor interface (BTI) and are closely linked to chemotherapy resistance in patients with brain metastases. According to the scRNA-seq results, vascular endothelial cells are mainly divided into three clusters (EC-1, EC-2 and EC-3), of which EC-3 is arterial endothelial cells (marker genes are EFNB2 and GJA5), EC-2 is venous endothelial cells (marker genes are AKR1 and NR2F2), and EC-1 is in between. EC-1 clusters express genes associated with processes such as angiogenesis and collagen deposition, while EC-2 clusters express genes associated with hypoxia, inflammation, and antigen presentation.
In addition, the researchers also detected six multispecific ATP binding cassettes (ABCs) in endothelial cells, including ABCB1 (i.e., multidrug-resistant protein 1, MDR) and ABCG2, both of which are key molecules in the normal blood-brain barrier and are significantly more expressed in the region where venous (EC-2 clusters) endothelial cells converge (Figure 3).
Figure 3 Six multispecific ATP binding cassettes (ABCs) were detected in endothelial cells, and ABCB1 and ABCG2 were significantly highly expressed in the area where venous (EC-2 clusters, red) endothelial cells converged
The researchers then analyzed the characteristics of infiltrating immune cells in brain metastases and found that the immune cells in brain metastases were dominated by T cells and macrophages, of which T cells exhibited higher heterogeneity and consisted mainly of 5 clusters: CD8+ effector memory T cells (T: CD8+: EM clusters, indicating CD8+ effector memory T cells, and so on), central memory T cells (T:CD4+:CM1 clusters, T:CD4+:CM2 clusters, and T: T:CD4+:CM2 clusters, and T: CM clusters) and regulatory T cells (Tregs clusters).
These cells have different immune states and functions, such as T:CD8+:EM clusters rich in cytotoxic molecules (such as GZMA and IFNG), T:CD4+:CM clusters represent an activation state of immunomodulatory markers such as LTB and IL32, while T:CM clusters expressing low levels of CD4 and CD8A also express LTB and IL32, but they are also enriched with interferon-related genes (such as ISG15 and MX1).
To better understand these T cell state and functional transitions, the researchers used a nonlinear dimensionality reduction technique to capture geometric structures and functional states in high-dimensional data [4]. This analysis showed that the five T cell clusters were arranged in an orderly manner by the fraction of diffusion component 1 (DC1).
Among the first 50 genes positively correlated with DC1, there are molecules associated with lymphocyte activation (such as CORO1A and LCK), molecules associated with chemotactic and migration (such as CCR7 and CXCR4), and the expression of these first 50 DC1-related genes has a strong correlation with the T cell activation gene cluster (GO: 0042110), indicating that different T cell activation states drive T cell diversity. (Figure 4)
Figure 4 The five T cell clusters are arranged in an orderly manner by diffusion component 1 (DC1) fraction, and there is a strong correlation between the expression of the first 50 DC1-related genes and the T cell activation gene cluster (GO:0042110).
To understand the role of each T cell state associated with the cellular microenvironment, the researchers evaluated the expression of genes that regulate the microenviron and metabolism, and found that interferon signaling-related genes predominate in the T:Cm cluster, the depletion characteristics associated with high T cell activation are enriched in the T:CD8+:EM and T:CD4+:CM clusters, while those with lower DC1 scores (T:CD4+:cm2, Treg, and T:Cm) exhibit strong non-functional features.
In addition, this phenotypic shift from activated T cells to non-functional T cells coincides with the shift in cell metabolism from glycolysis and tricarboxylic acid cycles to lipid metabolism, and the above results reveal that T cell functional states (from activation/depletion to non-functioning) are associated with concomitant metabolic and microenvironment reprogramming.
Finally, the researchers also studied tumor cells, the main body in brain metastases, and identified two potential subpopulations of tumor cells through non-negative matrix decomposition (NMFS)[5], and found the same phenomenon in external datasets and animal brain metastase models: a group of genes related to high-expression proliferation and pre-mRNA splicing, and another group of genes related to stress, inflammation, translation, and extracellular matrix deposition. The production of these two types of cell populations is likely to be associated with the extracellular matrix.
The study also has certain limitations, due to the lack of matching data on extracranial primary tumors, making it impossible to study the differences and links between brain metastases and primary tumors.
Overall, this study comprehensively analyzed the cell types and functional states in brain metastases, and through the analysis of single-cell resolution of brain metastase cells, immune cells and other stromal cells, it deepened the understanding of the interaction between the inherent characteristics of tumor cells and the characteristics of the cell microenvironment in tumor brain metastases, which contributed to the development of future therapies for brain metastases.
bibliography
1.Kamp MA, Slotty PJ, Cornelius JF, Steiger HJ, Rapp M, Sabel M: The impact of cerebral metastases growth pattern on neurosurgical treatment. Neurosurg Rev 2018, 41(1):77-86.
2.Achrol AS, Rennert RC, Anders C, Soffietti R, Ahluwalia MS, Nayak L, Peters S, Arvold ND, Harsh GR, Steeg PS et al: Brain metastases. Nat Rev Dis Primers 2019, 5(1):5.
3.Gonzalez H, Mei W, Robles I, Hagerling C, Allen BM, Hauge Okholm TL, Nanjaraj A, Verbeek T, Kalavacherla S, van Gogh M et al: Cellular architecture of human brain metastases. Cell 2022, 185(4):729-745 e720.
4.Haghverdi L, Buettner F, Theis FJ: Diffusion maps for high-dimensional single-cell analysis of differentiation data. Bioinformatics 2015, 31(18):2989-2998.
5.Puram SV, Tirosh I, Parikh AS, Patel AP, Yizhak K, Gillespie S, Rodman C, Luo CL, Mroz EA, Emerick KS et al: Single-Cell Transcriptomic Analysis of Primary and Metastatic Tumor Ecosystems in Head and Neck Cancer. Cell 2017, 171(7):1611-1624 e1624.
Responsible editor 丨IoTalker